Applied Animal Behaviour Science 155 (2014) 1227

Behavioral differences among breeds of domestic dogs (Canis lupus

familiaris): Current status of the science
Lindsay R. Mehrkam
a
,

, Clive D.L. Wynne

a
,
b

a
Department of Psychology, University of Florida, Gainesville, FL 32611, USA
b
Department of Psychology, Arizona State University, Tempe, AZ 85287, USA

Article history:
Accepted 17 March 2014
Available online 22 March 2014
a b s t r a c t
In both popular media and scientific literature, it is commonly stated that breeds of dogs differ behaviorally in
substantial, consistent and predictable ways. Since the mid-twentiethcentury, scientists have asked whether
meaningful behavioral differences exist betweenbreeds of dogs. Today, there are over 1000 identified dog breeds in
the world, but to date, fewer than one-quarter of these are represented in studies investigating breed-specific
behavioral differences. We review here scientific findings of breed differences in behaviorfrom a wide range of
methodologies with respect to both temperament traits and cognitive abilities to determine whether meaningful
differences in behavior between breeds have been established. Although there is convincing scientific evidence for
reliable differences between breeds and breed groups with respect to some behaviors (e.g., aggression, reactivity),
the majority of studies that have measured breed differences in behavior havereported meaningful within-breed
differences has well. These trends appear to be relatedto two main factors: the traits being assessed and the
methodology used to assess thosetraits. In addition, where evidence for breed differences in behavior has been
found, thereis mixed consistency between empirical findings and the recognized breed standard. We discuss both the
strengths and limitations of behavioral research in illuminating differences between dog breeds, highlight directions
for future research, and suggest the integration of input from other disciplines to further contribute to our
understanding of breed differencesin behavior.

2014 Elsevier B.V. All rights reserved.

Contents
1. Introduction
2. Origins of dog breeds
2.1. Defining the breed standard: The need for research
3. Temperament and personality
4. Aggression
4.1. Owner-directed aggression
4.2. Stranger-directed aggression
4.3. Dog-directed aggression
4.4. Summary of breed differences in aggression
5. Emotional reactivity
5.1. Excitability and general activity
5.2. Exploratory behaviors: Avoidance and
6. Trainability
7. Cognition
7.1. Physical problem-solving tasks
7.2. Human-responsiveness tasks
7.3. Conclusions on breed differences in cognition
8. Discussion
8.1. Evidence for behavioral differences between breeds versus within-breed
8.2. Evaluating scientific evidence for the breed standard and breed categories
9. Conclusions
Acknowledgments
References

Corresponding author at: Department of Psychology, University of Florida, 945 Center Dr., PO Box 112250,
Gainesville, FL 32611, USA. Tel.: +1 610 392 5938.
E-mail addresses: lrmehrk@gmail.com, mehrky14@ufl.edu (L.R. Mehrkam).
http://dx.doi.org/10.1016/j.applanim.2014.03.005
0168-1591/

2014 Elsevier B.V. All rights reserved.

1. Introduction
It is obvious that breed differences in behavior are bothreal and important in magnitude.Scott and Fuller (1965), p.
385. To many, the above quotation from Scott and Fullersclassic Genetics and the Social Behavior of the Dog (1965)
seems intuitively obvious. Over 1000 distinct dog breedsare in existence today (Morris, 2008), and of these,
approximately 20% are recognized by various nationaland international kennel clubs. Canis lupus familiaris is a
subspecies that exhibits extraordinary variation in morphological phenotype; it is assumed that breeds therefore differ
widely in their behavioral phenotypes as well. Although the number of recognized dog breeds has grown substantially
over the last century and a half, relatively little empirical research has targeted the behavioral characteristics that
define breeds. As a result, dog breeding has beenreferred to as a well-established art, but a crude, unestab-
lished science (Rine, 1991). Yet, it is widely acknowledged that different breeds have different and consistent
behavioral predispositions due to selective pressure by humans (Howell and Bennett, 2011; Bradley, 2011; Serpell
and Hsu, 2005). The domestic dog and the breeds that make up this diverse subspecies is becoming an
increasingly popular subject for behavioral research. The goal of this paperis to review and summarize research
findings that have measured differences between dog breeds in temperament and cognitive performance in order to
determine whether there is empirical evidence that dog breeds differ in substantial and systematic ways. We consider
the historical understandings of breed origin, the diverse methodologies that have been used in empirical literature,
their validity and appropriateness, the developmental history of subjects tested, and the extent to which specific
breeds havebeen represented in the scientific literature. We hypothesize that (1) studies that have compared
behavior between and within dog breeds suggest that within-breed differences are more evident than between-breed
differences, and (2) where evidence for breed differences is found, the breed standard will be concordant with the
empirical conclusions for the trait that was measured. Finally, we offerconclusions about assessing breed differences
in behaviorand identify areas of research that need further investigation or replication.

2. Origins of dog breeds

The first evidence of distinctive dog breeds in the histor-ical record appears around 3000 to 4000 years
ago (Breweret al., 2002; Harcourt, 1974). While many of the breed classes of dogs we know today, as
well as their qualities and traits, were well defined by the Roman period (Clutton-Brock, 1995; Xenophon,
1897), the major era for the proliferation of dog breeds in Europe was the Middle Ages (Clutton-Brock,
1995). The unparalleled phenotypic variability between breeds led even Charles Darwin to believe that the
domestic dog must have had at least two common ancestors (Darwin, 1859). In fact, it was during
Darwins lifetime that modern breeds became closed pop-ulations (Ritvo, 1987). Biologically, dog breeds
are groups of individuals within the subspecies C. familiaris that strongly resemble one another based on
a series of characteristics that are identifiably different from other groups (Brewer et al., 2002). However,
the most accurate way to define a breed basedon behavior is still debated to a great extent. One
hypothesis is that dog breeds originated as land races adapting to different environments in geographic
isolation, with little to no human selective breeding (Coppinger and Coppinger,2001). An alternative
hypothesis is that artificial selection by humans was the fundamental basis of establishing dog breeds,
accomplished by mating dogs with similar characteristics to produce similar future generations (Rine,
1991). The true origin of dog breeds is clearly a subject of considerable debate (Larson et al., 2012). This
uncertainty has also prompted questions about how exactly a breed should bedefined, and is in itself a
source of disagreement among scientists today.

2.1. Defining the breed standard: The need for research

Regardless of how breeds may have originated, the wide variation in the dog breeds we know today is
produced and maintained by ongoing human selection for specific mutations, and for variations in size,
coat type and color (Coppinger and Coppinger, 2001; Scott and Fuller, 1965). Despite the long history of
human selection for specific traits in dogs, ancient and modern breeds are not closely related because of
to a loss of genetic diversity due to both historical (e.g., World Wars), cultural and geographic factors
(Larson et al., 2012), and this has contributed to the debate of how to define a breed. The opposing
views on the proper definition of a dog breed are largely reflected in the two distinct ways in which breeds
are categorized. The more recent approach groups breeds according to their genetic similarity. Parker and
Ostrander (2005) used microsatellite DNA across the dog genome to genotype 85 breeds into four
clusters Asian/African breeds, mastiff-type dogs, herding dogs and sighthounds, and modern
hunting dogs with some breeds being assigned to more than one cluster. These clusters were identified
as genetically distinct subpopulations based on patterns of allele frequencies, suggesting that breeds of
dogs are local and temporal phenomena recently derived by crossbreeding, such that most modern breeds
of dogs remain closely related (Larson et al., 2012; Wayneand Ostrander, 2007). The more traditional and
widespread method of categorizing breeds has been promoted by various national and international
kennel club organizations, in which dog breeds are split into groups based on the nature of the task with
which the dogs are associated. While dozens of such organizations exist, empirical studies that have
examined breed differences in behavior have largely focused on the categorizations determined by the
American Kennel Club (AKC) (180 recognized breeds in seven breed groups) and the Fdration
Cynologique Internationale (FCI) (365 breeds, 10 breed groups) (see Table 1 for all groups and
descriptions). It is often claimed that some modern breeds were originally bred specifically for various
physical attributes that also led to changes in behavior (e.g., Toy or Companion breeds as defined by the
AKC), where as other breeds may have been selected specifically for behavioral traits that are useful for a
particular task and these traits subsequently came to characterize the different breeds of dogs (e.g.,
Herding breeds, livestock-guarding breeds) (American Kennel Club, 2012; Morris, 2008). It isworth noting
that there is considerable overlap between the FCI breed groupings and the genetic clusters outlinedby
Parker and Ostrander (2005) relative to the groups employed by the AKC (Table 2). Irrespective of
different categorizations, there appears to be widespread agreement that dog breeds differ in behavior
from one another in measurable and predictableways; in other words, purebred dogs have a specific
behavioral conformation, or breed standard (Coppinger and Coppinger, 2001). The breed standards for
153 recognized breeds are published by the AKC in The Complete Dog Book (2006) and on AKCs official
website. For example, the Golden Retriever is described as friendly, reliable, and trustworthy, while the
German Shepherd Dog (GSD) has a certain aloofness that does not lend itself to immediateand
indiscriminate friendships. Although this terminology may appear consistent with popular conceptions of
these breeds, such accounts are almost completely anecdotal. Wewill therefore examine not only the
empirical evidence of breed differences in temperament and cognition in this review, but also whether
these findings confirm or conflict with breed standards and other popular stereotypes.

3. Temperament and personality

The majority of empirical literature on breed differences in behavior deals with temperament. The term
temperament may refer to either individual or breed behavioral differences that emerge early in
development, are elicited in a range of situations, and remain relatively stable overtime (Diedrich and
Giffroy, 2006). Included in temperament are personality traits, although the terms personality and
temperament are often used interchangeably (Jones and Gosling, 2005; McCrae et al., 2000).
Although there is no universally agreed upon definition, personality are similarly defined as behavioral
characteristics of individualsthat are consistent and generalizable across settings and is the result of
interaction between temperament and the environment (e.g., Jones and Gosling, 2005). The study of
personality traits in dogs has gained a great deal of empirical attention in recent decades. Personality may
be assessed experimentally (i.e., a behavioral battery test, e.g., Jones and Gosling, 2005; Svartberg,
2002, 2007), observationally in a naturalistic setting (e.g., Goddard and Beilharz, 1983), or by indirect
assessment via an informant (e.g., giving an owner a questionnaire used to rate personality traits, as in
Jones and Gosling). Because we feel theterm personality has anthropomorphic connotations (e.g.,Jones
and Gosling, 2005), we will generally use the term temperament in this review, but will refer to
personality. In dogs, personality has been assessed, for example, by the traits aggressiveness, sociability,
curiosity/fearlessness, and playfulness (Svartberg and Forkman, 2002; Svartberg,2006) or along the
shyness-boldness spectrum (Svartberg,2002). Similarly, temperament has been defined by the broad
traits of aggressiveness, reactivity, and trainability (e.g., Hart and Hart, 1985; Takeuchi and Mori, 2006).
Hart and Miller (1985) conducted the largest study todate on breed differences in temperament in which
48 veterinarians and 48 obedience judges ranked seven breeds chosen randomly from 56 breeds of dogs
on 13 behavioral traits. A subsequent cluster analysis revealed that of these 13 traits, aggression,
emotional reactivity and trainability were found to account for the most variance between breeds (Draper,
1995; Hart and Hart, 1985). Indeed, many studies on breed differences in temperament in dogs have
investigated one or more of these three traits.

4. Aggression

Aggression is the most common and serious behavioral problem reported in dogs kept as companion
animals (Borchelt, 1983; Fatj et al., 2007). The effect of breed on the expression of canine aggression
has been suggested by numerous experimental and survey-based studies (Duffy et al., 2008; Hsu and
Sun, 2010; Scott and Fuller, 1965; Svartberg and Forkman, 2002). The measurement of aggression in
dogs employs a wide range of sources, including dog bite statistics, surveys of dog owners, survey and
referrals of general veterinary practitioners and databases from companion animal behaviorists (Duffy et
al., 2008; Fatj et al., 2007). Several topographies of aggression have been reported in a wide range of
breeds and breed groups, making aggression one of the most complex canine behaviors to define
in terms of context, intensity, and target.

4.1. Owner-directed aggression

Owner-directed aggression (ODA), or agonistic behavior that is directed toward people who are familiar to
the dog, represents the most common cause of dog bite injuries reported (Overall and Love, 2001).
Although no published direct assessments of breed differences in ODA exist, many indirect sources are
available. Incident-based reports, including statistics on bite-related fatalities andveterinary caseloads, as
well as owner-based questionnaires, yield inconsistent findings of breed differences in aggression. For
example, Sacks et al. (2000) found that Pit Bull-type dogs and Rottweilers were involved in more
than half of 238 total human deaths caused by a dog bite-related injury in the United States
between the years of 1979 and 1998, whereas breeds that have been anecdotally considered to
exhibit low ODA were ranked highest forbite-related fatalities where statewide breed-specific legislation is
practiced (University of Colorado Denver (UCD), 2010). The difficulties with drawing definitive
conclusionsfrom bite statistics alone have been acknowledged (Sackset al., 2000), as certain breeds and
breed groups may be either over- and under-represented in different biting populations (Cornelissen and
Hopster, 2010). Owner-based surveys using the standardized questionnaire C-BARQ
TM
in the United
States, Japan, and Taiwan have shown substantial discrepancies in breeds that exhibithigh ODA. The
highest levels of ODA have been reported in mostly small to medium-sized breeds (Duffy et al.,
2008; Hsu and Sun, 2010), and many of the same breeds ranked high for ODA by Japanese dog owners
also ranked high for snapping at children and exhibiting dominance over owner (Takeuchi and Mori,
2006). However, in a study by Hartand Miller (1985), breeds ranked by judges as showing the greatest
tendency to exert dominance over their owners or snap at children were not consistent with these more
recent owner surveys. Studies of veterinary case loads consistently report the English Cocker Spaniel
(Amat et al., 2009; Beaver, 1983; Borchelt, 1983; Fatj et al., 2007; Wright and Nesselrote,1987) and
the English Springer Spaniel (Guy et al., 2001; Landsberg, 1991; Reisner et al., 1994) as exhibiting
the highest levels of ODA. These trends in ODA between breeds obtained by indirect assessments may be
due to different perceptions of petdogs aggression between veterinarians and owners.

4.2. Stranger-directed aggression

There have been several direct behavioral assessments of breed-specific stranger-directed aggression
(SDA). Bollen and Horowitz (2008) assessed over 2000 shelter dogs and found that failure on a behavioral
evaluation (e.g., exhibiting serious aggression, including lunging while growling and snarling, and any
attempts to bite) was significantly higher for high risk (Pit Bull, Rottweiler, ChowChow, Husky) than
for low risk (all other) breeds. It should be noted that these results may be confounded by a testers
prior and potentially biased perceptionof the likelihood of these breeds to exhibit SDA. Breed groups
did not differ in their reactions when dogs on-leashoutdoors were approached by an unfamiliar person in
afriendly manner (Vs et al., 2005). On breed level, Belgian Shepherds (Tervuerens and Groenendaels)
exhibited significantly higher levels of avoidance behavior than both sled dogs (which included Malamutes
and Huskies) and retrievers (which included Golden Retrievers and Labrador Retrievers) when approached
by an unfamiliar human. The comparability of these experimental studies is limited, however, by
differential inclusion of breeds, dissimilar testing situations and the inconsistency of measures to quantify
SDA. Breeds ranked high for SDA were consistent across owner surveys using C-BARQ
TM
(Duffy et al.,
2008; Hsuand Sun, 2010), but were not consistent with experimental results. This is likely due to the
relatively low diversity of breeds included in experimental studies and the results are not readily
comparable between methodologies. Breeds that scored low on SDA on C-BARQ
TM
, however, were
consistent with the results reported by Vs et al. (2005). Relatively average scores were reported for SDA
on C-BARQ
TM
among breeds commonly targeted in the mediafor human-directed aggression (e.g., Pit
Bull-type breeds, Rottweiler, German Shepherd Dog). There is evidence that behavioral differences within
a breed may also be related to whether the individual dog of a particular breed has been bred for show or
field work. Breeding for show has generally been associated with lower levels of aggression, curiosity,
trainability (Serpelland Hsu, 2005) and playfulness, and with higher levels off earfulness, while selection
for field work was correlated with higher levels of playfulness and aggression (Svartberg,2006). There is
evidence of this pattern in ODA scores for Labrador Retrievers. The reverse effect was found for English
Springer Spaniels, such that conformation-bredindividuals bredindividuals tended to be more owner-
aggressive than field-bred individuals (Duffy et al., 2008).
Aggressive individuals within a breed have also been linked to yellow coat color in Labradors (in Houpt
and Willis, 2001; Kobelt et al., 2007) and golden coat color in English Cocker Spaniels (Amatet al., 2009;
Podberscek and Serpell, 1996). This may be due to an overlap in the biochemical synthesis pathways of
melanin and dopamine and other neurotransmitters (e.g., low levels of serotonin) that contribute to the
expression of aggressive behavior (Hemmer, 1990). Associations between coat color and aggression seem
breed-specificas no such effect was observed in Miniature Bull Terriers (Schalke et al., 2010) or Jindo
Dogs (Kim et al., 2010).

4.3. Dog-directed aggression

Scott and Fuller (1965) were the first to systematically observe, measure, and document clear breed
differencesin the tendency to exhibit intraspecific aggression. Their experiments suggested that breeds
defined as aggressive may be more likely to develop a linear dominance hierarchy.

Their findings are similar to todays AKC (2006 Guide) temperament standards for the five breeds tested
in their experiments (Basenji, Beagle, Cocker Spaniel, Shetland Sheepdog, and Wire-Haired Fox Terrier).
Importantly, Scott and Fuller reared several litters of each breed (with sample sizes ranging from 21 to 57
individuals for each breed) in controlled environments. In contrast, most of the empirical literature on
breed differences in dog-directed aggression (DDA) is based on surveys and indirect reports, usually from
owners, breed judges, veterinarians, and other animal care professionals. Generally consistent breed
differences in intraspecific aggression have been found using the C-BARQ
TM
in the United States, Japan
and Taiwan, with some exceptions. This is noteworthy, given both cultural differences in dog ownership,
as well as restricted gene flow between distant geographic locales. In addition, the severity of injury
resulting from intraspecific aggression has been attributed to breed-specific aggressiveness in anowner
survey study in Germany (Roll and Unshelm, 1997). Relatively few veterinary caseloads have been
published with respect to DDA, as most of these reports consist of dogs that exhibit human-directed
aggression.The tendency to exhibit behavioral precursors to DDA may be related to breed differences in
social signaling. Draper (1995) found intraspecific aggression to be positively correlated with the
uprightness of ear structure, whereas overall size was negatively correlated with reactivity. Overall
morphology seems unrelated to the sending or receiving of any social signals (Kerswell et al., 2010),
though some specific physical characteristics (e.g., snoutlength, reduced skull shape) have been
correlated with signaling frequency (Roberts et al., 2010). Short-snouted (brachycephalic) dogs in
particular may appear morejuvenile and less threatening, and thus more likely toelicit play behaviors from
other dogs. Such phenotypes may make it difficult for these breeds of dogs to communicate with other
dogs. Breed differences have been found in the frequency of tail-wagging (Scott and Fuller,1965), and
across breeds the frequencies of dominant and submissive behavior patterns are positively correlated with
physical resemblance to the gray wolf (Goodwinet al., 1997). However, even between somewhat similar
breeds, levels of conspecific interactions do not appear to be qualitatively different between Cocker
Spaniels, Labradors, and Miniature Schnauzers except for in the first 3 min of an encounter when Cocker
Spaniels showthe least interactions (Pullen et al., 2013). The influence of morphological differences on
social communication between breeds is in need of further research.

4.4. Summary of breed differences in aggression

Breed differences in aggression have been reported in numerous studies. Many breeds were differentially
ranked for DDA, ODA, and SDA, suggesting that environmental stimuli, rather than breed alone, play a
major role the propensity to exhibit aggression. Breed rankings for ODA are more consistent across
methodologies than either DDA or SDA. Across aggression topographies, however, breeds exhibit
differential levels across methodologies. Bite statistics are confounded in that bites from large dog
breedsare more likely to lead to hospitalization and fatalities than bites from smaller breeds. Veterinary
caseloads and owner surveys appear more consistent in breed rankings of aggression, but neither of these
indirect methods of assessing breed-specific aggression are in agreement with experimental data. Few
experimental studies have been conducted on aggression in dogs, with fewer different breeds studied
compared to survey-based studies. The disproportionate amount of indirect assessments relative to
experimental studies is likely due to ethical and practical difficulties with using aggression as a dependent
measure. However, if behavioral precursors to aggression can be identified accurately and reliably, it may
become possible to study aggression more frequently in controlled experimental studies. The low
reliability between direct and indirect methods for measuring aggression may also be attributed to
different perceptions of what constitutes aggression. The seperceptions may be influenced by experience
with dogs of different breeds, the number of dogs owned, the size and appearance of the dogs, and the
demographics of respondents. Across methodologies, however, assessments of breeds exhibiting low
aggression (e.g., Golden Retrievers, Labrador Retrievers) are relatively consistent. Within-breed
differences have been reported in direct, but not indirect assessments (e.g., owner surveys) of
aggression. Owner surveys were not consistent with popular perceptions of breeds as portrayed in breed
standards. All of these factors, however, further complicate the ability to draw conclusions about clear
breed differences in aggression.

5. Emotional reactivity

Scott and Fuller (1965) defined reactivity as theintensity of all responses, external and internal,
made to a stimulus change (p. 194). Emotional reactivity has been characterized by traits such as
excitability, general activity, snapping at children, excessive barking, demands for affection (Draper,
1995; Hart and Miller, 1985), and fearfulness or or distress (Goddard and Beilharz, 1985a). We consider
three major components of emotional reactivity here.

5.1. Excitability and general activity

Serpell and Hsu (2001) defined a dog high in excitability as, high [in] energy, [and having] a very
difficult time settling down and behaving quietly for extended periods of time. (p. 351). Scott
and Fuller found clear differences between breeds in their development of directly observable emotional
reactions to human beings during social interactions (such as tail wagging, and distress vocalizations).

Scott and Fuller measured behavioral and physiological indicators of emotional reactivity between breeds,
by presenting dogs with different situations while briefly isolated from humans and conspecifics. An
observerrated dogs behavioral responses on a 5-point intensity scale. Highly significant differences
between all five breeds were found for shock response, escape activity, erect body posture, tremor,
investigation, attention to observer, lip licking, vocalization, panting, tail wagging, resistance to forced
movement, and biting. Elimination, heart rate change during bell ringing, and handler effect score were
not significantly different between breeds. Individual differences within breeds were observed as well.

Indirect assessments have ranked the Miniature Schnauzer, Scottish Terrier, and West Highland White
Terrier highest for all traits (e.g., playfulness, destructiveness, excitability, and excessive barking) that
were considered to be indicators of reactivity (Hart and Miller, 1985; Notari and Goodwin, 2007). These
findings have been relatively consistent with owner reports in the United States and the United Kingdom
(Bradshaw and Goodwin, 1998). More recently, online owner reports have suggested significant
differences between breed groups in measures of in attention and hyperactivity-impulsivity, with hounds
havinghigher scores for in attention (also correlated with excessive barking, constant motion, and
difficulty maintaininga stay) than other breed groups (Lit et al., 2010), and Herding group breeds scoring
higher than Toy and Non-Sporting breeds for enjoying active play, constant motion and anticipation.
These findings suggest that substantial breed differences may exist along different dimensions of a single
trait.

Physiological and structural conformations may further substantiate differences in excitability between
breeds, though such cross-disciplinary research in this area is still relatively sparse.

Arons and Shoemaker (1992), for example, found differences in measured neurotransmitter levels
(norepinephrine, dopamine, and epinephrine) between Border Collies, Siberian Huskies, and Shar
Plaininetz (a livestock-guarding breed), which corresponded with excitability levels of these breeds (Arons
and Shoemaker,1992). Breed-specific dietary metabolism differences have also been found between
Labrador Retrievers and Miniature Schnauzers in an experimental comparison between metabolite
fingerprints generated from urine samples via flow injection electrospray (Beckmann et al., 2010). This
may represent an additional physiological basis for breed differences in excitability and other measures
ofactivity, as the ability to sustain high-intensity effort is directly related to maximum energy consumption
regulated through metabolic rate (Coppinger and Coppinger, 2001).

5.2. Exploratory behaviors:

Avoidance and investigation

Differences between breeds of dog in fearfulnessdefined as a tendency to withdraw from intense
and novel stimuli (Goddard and Beilharz, 1985b), have been reported in numerous studies in a variety
of contexts (Goddard and Beilharz, 1985b; Mahut, 1958; Plutchik,1971; Scott and Fuller, 1965). Novel
object and startle tests are commonly used to measure fear in dogs (Kinget al., 2003; Ley et al., 2007).
In such procedures, subjects are presented with potentially fear-evoking stimuli (e.g., door banging,
whistling, opening umbrellas) and the dogs responses are recorded. Fearful responses typically include
withdrawal, crouched posture, or lowered tail position (Goddard and Beilharz, 1985a). Such behavioral
tests are the most common in assessing temperament traits, including fearfulness, in dogs (Jones and
Gosling, 2005).

Mahut (1958) experimentally determined significantbreed differences in response to startle stimuli. Dogs
of 10 different breeds were either kept as pets, kennel-reared (given access to the outdoors and other
conspecifics, but not treated as pets), or cage-restricted (reared in isolation from humans and
conspecifics). Kennel-reared dogs showed more avoidance responses than the other two groups.

Similarly, Beerda et al. (2000) reported Beagles housed individually in kennels exhibited more oral
behaviors and sighing after they were startled by a slamming door, that the authors termed breed-
specific stress-behaviors (p. 59). However, this effect was not observed in privately-owned Beagles or
pair-housed kenneled Beagles, suggesting that within-breed differences may be due to learning of
differential consequences based on dissimilar housing environments. For example, home-reared
individuals may have more experience with learning that the consequences of a slamming door (e.g., the
arrival or removal of a person) than would kennel-reared individuals. This suggests an environmental or
ontogenic influence on reactivity rather than a genetic one.

Similarly, Mahut (1958) also observed significant within-breed variation, yet concluded that there are
clear breed differences in susceptibility to fear patterns. More recently, support for this claim has been
found in that breeds bred for rat hunting and fighting exhibit less fearful responses to aversive stimuli
than Sporting, and Herding breeds (Serpell, 1995), although it is important to note that the historical
function of many breeds remains unproven. Another recent survey study reported that Herding breeds,
Labrador Retrievers, Cocker Spaniels, English Springer Spaniels, and Toy Poodles, were at a significantly
decreased risk for exhibiting fearful responses toward loud noises (e.g., fireworks, thunder, gunshots)
(Blackwell et al., 2013). Breed differences in emotional reactivity are especially important to the working
and service dog industries. Dogs selected for guide dog work are typically high in confidenceand low in
aggression-dominance, whereas fearfulnessis the most common reason dogs are rejected (Goddardand
Beilharz, 1985a).

Labradors were found to be the least fearful in startling tests between 1 and 18 months of age, while
German Shepherds were the most fearful (Goddard and Beilharz, 1985a). These experimental findings are
consistent with breed differences in success rates for graduating as Seeing Eye guide dogs (Ennik et al.,
2006), where Labrador Retrievers and Labrador-Golden Retriever crosses were the most likely to succeed
in training whereas German Shepherd Dogs were the least likely. Furthermore, the topography of
fearfulness, rather than the likelihood of its expression, may differ between breeds. For example, Kelpies
and Boxers both displayed intermediate levels off earfulness during startle tests; however, Boxers
respondedby becoming inhibited, whereas Kelpies responded with high activity (Goddard and Beilharz,
1985a). Breeds reported as showing low fearfulness toward startle stimuli have been more frequently
reported as potential sheep herding dogs when compared to other breeds (Christiansen et al., 2001).

Green and Woodruff (1983) found that 11% of Komodorok and 78% of Great Pyrenees were successful in
deterring predation in rangeland flocks of sheep, and observed that Great Pyrenees were not as playful
with the sheep, less sociable to people and thus more apt to stay with the sheep rather than following the
shepherd and adapted more readily to new people and terrain, whereas Komdorok were more reactive
and immature. The relationship between working performance and response to startle stimuli has also
been reported in racing dogs. King et al. (2003) found that Greyhounds with such histories displayed
much less avoidance and aversion to novel stimuli than other breeds. Thus, breed differences in task
performances by working and service dogs may be further explained by their tendency to exhibit
neophobia or other types of reactivity to novel stimuli. There is some evidence for breed differences in
socialinvestigation during human-dog play in experimental studies (Scott and Fuller, 1965) and sociability.

Several experimental studies have found that Labrador Retrievers and Golden Retrievers are among the
breeds that score higher in sociability and curiosity (Svartberg, 2006; Vset al., 2005; Wilsson and
Sundgren, 1997), where as German Shepherd Dogs (Wilsson and Sundgren, 1997), and Poodles and
Belgian Shepherds (Vs et al., 2005) were often among breeds with the lowest scores. Golden Retrievers
have been reported to exhibit more indicators of stress (increased proximity-seeking behavior, locomotion
and passive behavior, and decreased play and exploration) than Labrador Retrievers in experimental
conditions in which the dogs were repeatedly separated from their owners (Fallani et al., 2007). This is
likely related to the higher rankings for Labrador Retrievers than Golden Retrievers in startle or emotional
reactivity tests and may furthermore have applications for trainability in service dog work. A longitudinal
study in the United Kingdom found no breed differences in the prevalence of separation anxiety between
Labrador Retrievers and Border Collies basedon owner responses (Bradshaw et al., 2002), where both
breeds are considered highly trainable as search and rescue dogs (Rooney and Bradshaw, 2004). The
personality dimension of shyness-boldness has been correlated with performance in working dogs, such
that bolder dogs perform better during working dog trials; though this has also been linked to the
experience of the handlers (Svartberg, 2002).

Svartberg (2002) observed that German Shepherd Dogs tended to perform better than Belgian Tervurens,
but detected a threshold effect amongs thigh-performing dogs, such that no breed or sex differences were
found. Breed-type had no predictive validity on working dog certification among German Shepherd Dogs,
Belgian Malinois, and Dutch Shepherds (Sinn et al.,2010).

Svartberg (2006) collected scores on a standardized test of responsiveness to diverse social and non-
social stimuli of 31 breeds in five groups (Working, Terriers, Herding, Hounds, and Gundogs), and
proposed that breed differences in fearfulness may be a consequence of recent selection for current breed
stock, rather than of past selection in breed origin. Within breed differences in reactivity have also been
reported. In Swedish populations, selection of a breed for use in dog shows correlates positively with
fearfulness and negatively with playfulness, aggressiveness, and curiosity; in contrast, breeding for
working dog trials correlates positively with playfulness and aggres-iveness (Svartberg, 2006).

6. Trainability

One of the most practical interests of dog ownersregarding different breeds of dogs is their trainability
alsoreferred to as obedience or working intelligence (Coren,1995). Trainability may be defined as a
dogs willingnessto attend to its owner and obey commands, with highmotivation and resistance to
correction, and low levelsof distractibility (Serpell and Hsu, 2005). Although thisarea appears to have
received less empirical attentionthan either aggression or emotional reactivity, trainabilityhas been
measured in different dog breeds with respectto quieting on command, leash training, housebreaking,
obedience training, and habit formation in both experi-mental (Scott and Fuller, 1965) and survey (Hsu
and Sun,2010; Hart and Hart, 1985) studies. In addition, the abil-ity to predict behavioral tendencies of
different breeds isconsidered a powerful tool for cost-effective training ofdogs for working dogs to be used
in a variety of roles(Riva et al., 2012). Experimental studies have found breeddifferences in performance
during leash training, quiet-ing, and obedience (specifically, staying in one place oncommand), (Scott and
Fuller, 1965). Breed differences intrainability have also been detected indirectly with ownerreports and
surveys (Serpell and Hsu, 2005). Extraversionand training focus were the only characteristics of train-
ability that were significantly different among the sevenbreed groups recognized by the Australian
National KennelCouncil (Ley et al., 2009). Herding breeds were report-edly the easiest to train, based on
a cluster analysis ofdata obtained from an owner survey (Turcsn et al., 2011).Traits of trainability and
boldness differed significantlybetween conventional and genetic groups, though behav-ioral breed clusters
showed poor correspondence to bothtraditional and genetic categorization. Breed rankings ofworking
intelligence are highly consistent with the resultsof other studies that have found breed differences in
obedi-ence training through survey responses by dog owners inboth the United States (Hart and Miller,
1985) and Japan(Takeuchi and Mori, 2006) and judges rankings in theUnited States (Coren,
1995).Within-breed differences in trainability have beenobserved as a function of rearing styles. For
example,Freedman (1958) tested puppies from eight litters andfound within-breed differences in
responses to reprimandsin Beagles and Fox Terriers that were raised in indul-gent environments (in
which all activities the puppiesengaged in were encouraged and they never receivedpunishment)
compared to disciplined environments (inwhich the puppies were restrained in the experimenterslap
and trained (p. 585). There is a large disparity in physi-cal differences among breeds and breed groups
which mayfactor greatly into trainability differences among dogs. Ina reanalysis of the data obtained by
Hart and Hart (1985),taller dogs tended to rate higher for trainability than didshorter breeds (Draper,
1995), and breeds perceived ashighly trainable according to Corens (1995) rankings werecloser to the
mean of all breed height standards sampledthan breeds considered to exhibit the worst trainabil-ity,
which included breeds at extreme ends of the heightspectrum (Helton, 2010). Thus, differences in
physical mor-phologies, rather than historical function, may allow somebreeds greater advantages at
some training tasks relativeto others. Neopohobia was found to influence trainabil-ity in dog breeds as
well. Furthermore, existing studies ontrainability are difficult to interpret because they include adiverse
range of methodologies, dependent measures (e.g.,rate, latency, performance criterion), and restrictive
breed selection.

7. Cognition

Where as the use of indirect assessments (e.g., questionnaires) is relatively popular for measuring
temperament and personality, experimental studies are more commonly used in assessing breed
differences in cognition. Although experiments on breed differences in cognitive abilities originated with
Scott and Fuller (1965), this area has received less empirical attention than temperament. However, the
scientific inquiry of canine cognition has been steadily developing over the past decade with a new focus
sensitivity to human cues. Thus, the literature on cognitive tasks in domestic dogs can be explored in two
contexts: tasks that involve non-human stimuli and tasks that involve human interaction or response to
human cues.

7.1. Physical problem-solving tasks

Much of the classic studies on both simple and complex learning in domestic dogs originated with physical
problem-solving tasks. In a manipulation task conducted by Scott and Fuller (1965), Basenjis were
consistently the most successful of the five breeds in removing a dish offood from within a covered
wooden box that could only bereached by nosing or pulling the dish out through an openside, indicating
that Basenjis are more skillful in manipulating objects. Cocker Spaniels consistently performed the worst,
whereas Beagles, Wire-Haired Fox Terriers, and Shetland Sheepdogs were all intermediate in
performance. All breeds improved with the presence of added manipulanda to make the reward more
accessible.

More recently, studies have explored the role of morphological differences in accounting for breed
differences in motor capabilities. McGreevy et al. (2010) found that short-snouted (brachycephalic) breeds
(e.g., Pugs and Boxers) took less time than long-snouted (dolichocephalic) breeds (e.g., Whippets and
Greyhounds) to complete a 100-paw criterion on a task of removing food from a cylindrical rubber Kong
toy. This suggests that short-snouted dogs use their paws more in manipulation tasks, whereas dolicho-
cephalic breeds may rely less on their paws because their relatively long muzzles allowed them to better
extract food from the Kong toys. This was not found with Shetland Sheepdogs or Fox Terriers in Scott and
Fullers (1965) manipulation task, indicating that morphology is not in itself sufficient to explain
differences in performance in manipulation tasks.

Since it is unlikely that rearing histories played a large role in the differences between these two
populations (puppies in both studies were under 16 weeks of age and attained from breeders),
discrepancies may rather reflect breed differences in for example thresholds of stimulation.

Breed differences in spatial perception have been assessed using a detour task (Scott and Fuller, 1965),
in which subjects had to move away from a visible dish containing food and walk around to the far side of
a transparent barrier to attain the food. Breed differences in maze performance were found by Scott and
Fuller in both performance, persistence (measured as number of attempts to get food) and improvement
measures, though background variance was highly influential in average error and time scores (Elliot and
Scott, 1965).

It was concluded that breed differences in performance were largely due to differences in fear responses
toward the apparatus between breeds (e.g., Basenjis exhibited the most fear responses, where Fox
Terriers exhibited the fewest). Wide individual differences were observed. More recently, Head et al.
(1997) found that the effect of age on open field activity is breed-dependent. In addition, exploratory
behavior measured by locomotor activity in an open field correlated with behavioral indicators of
cognitive function, measured by reversal learning, spatial learning, object recognition, and memory
acquisition.

More recently, differences in object permanence the ability to locate an object that is no longer visible
were not found between breed groups (terriers, sporting, and working) (Gagnon and Dor, 1992). Rather,
subjects were highly successful on visible displacement tests, but less successful on solving invisible
displacement problems, suggesting that dogs solved these tasks on the basis of visual information.

There are several possible explanations for the relative success of different breeds in physical problem
solving tasks. First, there may be breed-specific differential responses to frustrating or fearful situations.

For example, Scott and Fuller (1965) reported that Cocker Spaniels would simply lie down and become
inactive when they could not accomplish the detour task, whereas Basenjis remained active and were thus
more likely to solve the problem by chance. In the maze test, the success of Beagles was attributed to
their tendency to continuously investigate their surroundings, whereas Shetland Sheepdogs appeared
timid and hesitant in the maze and developed strong stereotyped habits a trait they attributed to the
central nervous system rather than environmental variables (Elliot and Scott, 1965).

Morphological differences may also be a factor. For instance, the cranial morphology of some
brachiocephalic dog breeds, (e.g., Pugs, Bulldogs) could affect a dogs sense of smell due to greater
forward rotation of their brains and consequently lower repositioning of the olfactory lobe (Roberts et al.,
2010), and thus influence their performance on odor or trailing tasks (Scott and Fuller, 1965), resulting in
differential reliance on smell as a function of breed. The number of possible variables strongly suggests
that understanding the causal factors for differences between breeds in physical problem solving is a
complex task in itself.

7.2. Human-responsiveness tasks

Breed-specific differences in responsiveness to humans were documented beginning with Freedmans
(1958) finding that punishment had a differential effect on social inhibition of eating in four breeds of
dogs. Using eight litters (two litters per each breed), observers reported that Beagles and Fox Terriers
were strongly oriented to the experimenter, whereas Basenjis showed increased activity levels but ignored
the experimenter in favor of inanimate objects. In contrast, Shetland Sheepdogs exhibited avoidance due
to fearfulness and exhibited lower activity levels (Freedman, 1958), and this is consistent with findings
that Shetland Sheepdogs performed poorly on tests in which their emotional reactivity interfered with
learning (Stafford (1996); Werboff et al., 1969; Takeuchiet al. (2001)). Recently, a number of studies
have examined domestic dogs ability to follow human cues to hidden food.

Pongracz et al. (2005) did not find differences between breeds of dogs in their ability to follow a human
point leading to a detour around a barrier.

A metaanalysis compared 24 breeds grouped by both the eight AKC categories and the four genetic
clusters identified by Parker and Ostrander (2005) used in 14 studies on a range of different pointing
types, and found no differences between breeds in their ability to follow a human point to hidden food
(Dorey et al., 2009). Wobber et al. (2009) found that working breeds (which included German Shepherds,
Belgian Shepherds and Siberian Huskies) performed better in following human cues to food (including
pointing and gazing) than non-working breeds (which included Toy Poodles and Basenjis), and the ability
the follow such cues was not influenced by breeds respective similarity to the graywolf. Other studies
have failed to observe differences in performance between herding and hunting dogs (Riedelet al., 2008)
and between gun dogs (Labrador Retriever, Golden Retriever, Cocker Spaniel) and non-gundogs (German
Shepherd Dogs, Terrier mixes, and Poodles) (McKinleyand Sambrook, 2000) in similar tasks.

Most recently, Udellet al. (2014) reported that inhibition of predatory motor patterns predicts faster
acquisition and higher levels of response accuracy on these tasks in breeds with exaggerated predatory
repertoires (e.g., Border Collies) compared to breeds that do not (e.g., Airedale Terriers, Anatolian
Shepherds). Given Scott and Fullers findings that Basenjis tended to be inattentive to the experimenter
during social interaction following in accessibility of food, it may be possible that Wobber et al.s findings
were due to behavioral idiosyncrasies of the particular breeds selected, rather than a characteristic of
non-working breeds in general, especially given the relatively small likelihood that the one or two breeds
chosen would be an accurate representation of an entire breed categorization.

Comparing breeds that are believed to differ in their physical traitsincluding size (Helton and Helton,
2010) and motor selections, rather than historical roles, may be a more fruitful approach.

Human-directed gazing has also been an active area of canine cognition research. Most recently, breed
differences in the acquisition of gazing at a persons face were not found (Jakovcevic et al., 2010), but
Labrador Retrievers and Golden Retrievers gazed at the human face significantly more during extinction
trials than did German Shepherd Dogs and Poodles. One explanation maybe that cooperative breeds
selected to work closely with humans are more skillful in cognitive tasks involving people than
independent breeds and mixedbreeds (Gcsiet al., 2009). Svartberg (2006) argued that selection during
recent decades may have led to great changes in breeds relative to their historical breed-typical behavior.

Thus, breeds may show differences in their communicative skills not only because of different histories of
selection, but also because their ability to learn such skills has been modified recently. Thus, the
differential rates of extinction in gazing may be a reflection of training experience during a dogs lifetime,
rather than phylogenetic influences. Passalacquaet al. (2011) found that Herding/Hunting breeds
exhibited more human-directed gazing during an unsolvable taskparadigm, than Primitive or Molossoid
breeds, but suggested this was influenced by age and thus experience with interspecific social
communication.

Furthermore, training in agility and search-and-rescue tasks alters dogs behavior on human
responsiveness tasks (Marshall-Pescini et al.,2009). Another consideration is that social reinforcers maybe
particularly salient to certain breeds or breed groups, so that gazing could act as a conditioned reinforcer
for one, but not another breed or breed group. A valuable area for future research would therefore be to
investigate how breeds differ in their motivation for food, social, or tangible (e.g., toy) reinforcers, and
how these differences influence performance on cognitive tasks.

7.3. Conclusions on breed differences in cognition

Many studies that have examined cognition in dogs have identified differences in performance among
various breeds, though this has been more common in tasks involving sensitivity to human gestures
rather than physical problem solving tasks. Where evidence in breed differences is found in the latter,
authors have attributed their findings to breed differences in morphology, whereas in the former breed
differences in sensitivity and propensity to follow human cues is attributed to artificial selection in certain
breed or breed groups linked to their historical roles. This latter argument should be made with caution,
as the intensity of selection for practical purposes has been fading as the importance of dog
companionship and breeding for appearance has risen, particularly in Western societies (Svartberg,
2006). Additionally, the lack of breed diversity in cognitive task studies presents another difficulty for
making broad inferences about the cognitive abilities of certain breeds and breed groups relative to
others. Of the 58 dogs in studies reviewed by Dorey et al. (2009), 19 were retriever or retriever mixes
and 17 were German Shepherd Dogs or mixes, followed by 5 Dachshunds, with no other breed
represented by more than three subjects. In addition, individual variation is an important consideration in
comparing performance on cognitive tasks, and many authors do not report individual data.

The discrepancies between breeds in many studies reviewed in this section may be due to ontogenetic
differences between individual dogs, rather than to genetic breed differences. Ontogeny in addition to
phylogeny plays a critical role in the ability of domestic dogs to respond to human gestures (Udell and
Wynne,2010). If studies aim to demonstrate breed-related differences in cognitive abilities, rearing and
housing conditions for subjects must be as similar as possible, as these factors can have a major impact
on at least some cognitive tasks (Dorey et al., 2009; Head et al., 1997; Udell and Wynne,2010). Test
situations should also be as standardized aspossible.

8. Discussion

Does the current literature provide evidence for breed differences in behavior, as Scott and Fuller stated
sosurely in 1965? Our objective here has been to provide a comprehensive review of the many different
sources of information available that claim evidence of breed differences in behavior, and to evaluate the
validity of each of these sources. Specifically, we tested two hypotheses: that there is more evidence for
within-breed variation in behavior rather than between breeds; and also, that where evidence for breed
differences exist, the scientific findings will correspond to the breed standard for a given breed. We
conclude that the question of breed differences in behavior is still very ripe for further investigation, and
would most greatly benefit from an emphasis of certain themes, which we will discuss here.

8.1. Evidence for behavioral differences between breeds versus within-breed
differences

Breed differences are more likely to be reported in survey-based studies and other indirect assessment
methods than in experimental studies, and in certain temperament traits rather than cognitive abilities
(Fig. 1). In studies measuring temperament, aggression and trainability are commonly assessed via
surveys and other indirect methods, whereas most of the direct assessments of temperament in dogs
have involved startle tests of emotional reactivity. However, the traits determined to have been
representative of aggression, emotional reactivity, and trainability may not be ideal variables for
delineating behavioral differences between breeds. Such traits may benefit not only from more
operational definitions, but maynot be entirely independent from other traits included inthese studies.


Fig. 1. Distribution of empirical research articles reporting breed
differ-ences and within-breed differences according to study type.

Survey-based studies may be extremely helpful in gaining information about breed-differences in
behavior, largely because it allows researchers to sample from awide range of breeds simultaneously.

However, survey-based studies are also difficult to interpret for several reasons. First, there is a lack of
base line data on the relative numbers of dogs of different breeds in the sample populations, making
comparisons of and representativeness less straigh tforward. In addition, respondents of surveys (which
typically tend to be veterinarians, owners, judges, or other individuals with extensive dog experience)
may be biased in their responses or may fail to accurately report the behavior of the dogs they interact
with on a daily basis. The validity of surveys and owner reports is a contentious issue: some argue that
these assessments are unavoidably subjective (Duffy et al., 2008; Nederhof, 1985), while others contend
that surveys are an acceptable method for assessing dog behavior (Block, 1961; Jones and Gosling,
2005).

Miklsi (2007) has perhaps stated most articulately and accurately the relationship between the two
methodologies: that while the use of questionnaires has been fashionable (p. 33) for collecting data on
behavioral comparisons between breeds, this method should notbe considered a substitute for research in
which the behavior of the subjects is the dependent measure.

Much work has already been done to develop robust, standardized questionnaires (e.g., C-BARQ
TM
) for
this purpose, but this research area is still relatively young. Many of the studies reviewed here have
reported not only substantial differences in behavior among breeds, but within breeds as well. Within-
breed differences are likely related to environmental differences in rearing histories (e.g., Freedman,
1958) or certain contingencies that the individual has contacted throughout its lifetime. The fact that
differences in temperament traits within breeds were also often reported, suggests the influence of either
genetic isolation within breeds across different geographic locations or environmental factors, such as
differential rearing conditions.

For example, Mirk et al. (2012) devised an individualized, adjective-based personality questionnaire for
dog owners and trainers in Hungary, and compared stranger-directed sociability, activity, aggressiveness,
and trainability) between the Hungarian Vizsla and the German Shepherd Dogand found no difference
between the two breeds after individuals were matched for demographic variables. In certain cases,
systematic within-breed differences in temperament have been observed as a function of whether the dog
was bred for show or field work.

Currently, the published sources that have reported breed differences in temperament traits have been
largely quantified by empirical, yet indirect methods. We believe more experimental research needs to be
conducted in this field, in order to address the inconsistencies reported among the more popular owner-
based survey studies.

Comparison of empirical findings across studies is greatly hindered by the different rearing environments
and demographics of the individuals within breeds being compared. Although the need to consider the
interaction between ontogeny and phylogeny when making conclusions about canid social behavior is not
a novel point (e.g., Coppinger and Coppinger, 2001; Fox, 1964; Scottand Fuller, 1965; Udell and Wynne,
2010) it appears to need considerable emphasis with respect to studies of breed differences in behavior.

In contrast, much of the research reviewed here suggests that there has been a bias in scientific studies
toward purely phylogenetic explanations for breed differences in behavior (Goodwin et al.,1997;
Svartberg, 2006; Wobber et al., 2009). An example of this would be attributing performance to selective
pressures during domestication. Yet, even in the most carefully controlled studies by Scott and Fuller
nearly 50 years ago, individual differences between breeds werefound (Cattell et al., 1973; Scott and
Fuller, 1965), and it has since been noted that some behavioral traits maybe more strongly influenced by
genetics (i.e., excitability, aggression) whereas other traits (i.e., housebreaking ease) appear to be more
influenced by environmental conditions (Hart and Miller, 1985).

In modern studies, commonly carried out on pets reared in relatively uncontrolled environments, we
expect environmental influences to have an even greater effect than in the older studies on kennel-reared
dogs. In addition, in consistencies in the results of many studies are not only due to different
methodologies and rearing histories of the subjects, but are likely also the result of differential inclusion of
breeds between studies. This is particularly true of the breeds used in Scottand Fullers experiments,
relative to the breeds examined in more recent experimental and survey studies. Certainly, the number of
breeds in existence today makes any sort of definitive and comprehensive examination of breed
differences in behavior daunting and many studies have therefore only examined the most readily
accessible and popular breeds. However, to seek differences among breeds in this way is shortsighted.

One major problem with the literature claiming breed differences in behavior is that many breeds are
grossly overrepresented and other breeds are virtually ignored. For example, although 202 breeds have
been used in the studies we have cited in this review, Labrador Retrievers, German Shepherd Dogs, and
Golden Retrievers were the most common breeds recruited (i.e., included in 30 or more scientific studies)
(Fig. 2).

In addition, this trend varies depending on the type of study being conducted, as has been acknowledged
in reviews of studies of breeds responding to human gestures (Dorey et al., 2009), and temperament
tests (Jones and Gosling, 2005). It is therefore premature and potentially misleading to study only readily
accessible breeds of dogs and use these findings to make generalities about the domestic dog as a
subspecies.

The fact that only small and unrepresentative samples of different types of breeds have been tested, adds
further emphasis to the need for caution when claiming breeds differ in systematic and fundamental ways.

Future research should thus place an emphasis on investigating behavioral traits of breeds that are
relatively under-represented in the literature. The lack of operationalized dependent measures with which
to assess behavior poses another difficulty for interpreting claims of breed differences in behavior. This is
shown by the higher consistency in breed rankings from studies that measure highly operationally defined
behaviors such as aggression and excitability compared to studies that attempt to measure less
operationalized concepts such as trainability, intelligence, or impulsivity. As a result, replications of
studies may be difficult to conduct and comparisons across studies may be more difficult if directly
observable behavior is not among the primary measures being assessed, and the use of anthropomorphic
terminology may hinder attempts to effectively explain why one breed is more likely to behave in one way
compared to another (Wynne, 2007).

Of particular interest maybe cases in which certain breeds have been observed to exhibit relatively low
food-motivation, because this maybe erroneously interpreted as poor obedience training or trainability. In
addition, there are also popular claims of breed-specific performance during particular tasks or
competitions, which have yet to be explored by experimental tests. Before claims of breed differences in
performance on various tasks can be made, it must be ensured that the reinforcers used for each breed
are indeed effective for that breed. More research clearly needs to be conducted in this area, as this is of
direct relevance for assessing the trainability of breeds. Another terminological concern in studying breed
differences in behavior is the practice of generalizing behavioral traits into excessively broad behavioral
categories. This likely leads to inconsistencies between studies, as behaviors that are only superficially
related to each other are being considered as if they represent the same class of behavior and have
related underlying causes.

Defininga ggression by structure or topography may not necessarily reflect the underlying behavioral
cause of the aggression (i.e., aggression due to negative reinforcement often termed fear-based
aggression compared to aggression due to positive reinforcement often termed dominance
aggression). In this case, we might argue that territorial aggressionis maintained by negative
reinforcement (the removal of an aversive stimulus contingent on aggressive responses, increasing the
probability that aggressive responses will be exhibited in the future) (Skinner, 1938). Surprisingly, very
few studies have examined breeds responses to particular types of reinforcement, especially in cognitive
tasks in which a breeds performance is the dependent


Fig. 2. Frequency of breeds represented in 20% or more of empirical behavioral studies. For veterinary
caseloads, only breeds reported were included. Othernames of breeds are denoted in parentheses in
cases in which the same breed was referred to by two or more different names across studies.

measure. Thus, describing breed differences in aggressionin terms of function rather than topography not
only eliminates the need to ascribe multiple levels of aggression, but also makes a functional treatment
less cryptic. Therefore, assessing how breeds differ in their propensity to exhibit behaviors based on their
function would be a constructive area for future research on breed differences in temperament. Similarly,
the general term intelligence is perhaps breed-specific but not absolute;
no breed is more or less intelligent in any general sense, but rather breeds differ in what they
have a predisposition to learn (Coppinger andCoppinger, 2001).
This is also true of measuring trainability and performance in cognitive tasksas any individual dogs
performance tends to be specific to particular test situations and is based on a large variety of capacities.
Attempting to rank dog breeds on a single dimension of intelligence thus seems an overly simplistic
method o fdetermining the relative behavioral performance of breeds of dogs.

A major argument against such rankings is that the intelligence of diverse breeds such as hunting or
herding dogs cannot readily be compared, as their skills differ qualitatively, not quantitatively. It is widely
acknowledged that any individuals propensity to exhibit a certain behavior can be influenced by genetic
selection (Bradley, 2011), and this can occur not only between breeds, but within breeds as well. It has
been proposed that behavioral genetic research can be useful to understand the genetic bases of different
traits, including temperament and personality (Saetre et al., 2006) and cognitive abilities (Dorey et al.,
2009), but morphological differences among breeds may also inform those with an interest in behavioral
differences among breeds. Such research could effectively translate these morphological differences into
behavioral differences, bridging the gap between two scientific disciplines with a common interest. It may
be the case that life experiences are highly important to all dogs, but that their behavioral tendencies are
a reflection of biological predispositions, both as a result of evolutionary history and morphology, aswell
as breed. Behavioral research on breed differences in dogs may therefore be better suited to examine
functional relations between a behavior of interest and environmental, rather than phylogenetic, variables.

Given the effects of the environment on within-breed differences in dogs, behavioral assessments and
treatments that address the environmental function of problem behaviors may prove especially useful.

8.2. Evaluating scientific evidence for the breed standardand breed categories

Based on our review of existing behavioral research, itappears that contrary to our hypothesis breed
standards are largely unsubstantiated, for most breeds that have been studied. This is concerning,
considering that breed is often used as a predictor of behavior when selecting dogs for working, or service
roles, and even companion animals (Bradley, 2011), and schemas or negative stereotypes of particular
breeds result in reduced likelihood of adoption from shelters (Wright et al., 2007). For instance, evidence
from owner reports and veterinary referrals contradict the merry and affectionate breed standard of the
English Cocker Spaniel, which should exhibit an even disposition (American Kennel Club, 2012, AKC
Meet the Breeds:English Cocker Spaniel, para. 3). There are some exceptions. For example, Golden
Retrievers have been found to exhibit relatively low aggression in numerous studies with a range of
methodologies, which is consistent with the friendly temperament noted by the American Kennel Club.

Definitive conclusions are prevented by the lack of diversity inbreeds that have been used in experimental
psychological studies, and these should be emphasized over indirect assessments, to provide some
measure of cross-validationfor the disproportionately large amount of surveys, referrals, and case studies
that have examined breed differencesin behavior.

A review of the literature on breed differences in behavior also suggests that both traditional and genetic
methods of grouping breeds may not be validated by behavioral research, especially with respect to
emotional reactivity and cognitive ability. The use of readily observable breed-typical behaviors as a
means of defining breeds, or categorizing similar breeds may offer advantages, as some breeds or breed
groups exhibit characteristic behavioral conformations that differ distinctively from other breed groups
and are assumed to be predatory motor patterns modified under domestication (Coppinger and
Coppinger, 1996).

These predatory motor patterns include orient, eye, stalk, chase, grab-bite, kill-bite, dissect, and
consume, each of which may be hypertrophied, ritualized, or suppressed. For example, livestock guarding
dogs must exhibit highly suppressed predatory behavior in order to perform their task without posing a
threat to livestock, and thus, rarely exhibit orient, eye, stalk, chase, grab-bite, or kill-bite (Coppinger and
Coppinger, 1996). In contrast, hounds have hypertrophied orient, mark (as opposed to stalk), chase,
grab-bite, and kill-bite, but lack eye, whereas retrievers exhibit hypertrophied orient and grab-bite but are
at fault for kill-bite. Herders possess hypertrophied orient, eye, stalk and chase, but suppressed grab-bite
and are faulted for kill-bite.

Thus, these readily observable breed-typical behavioral conformations still have ties to the historical roles
of breeds, but are more objective and directly exhibited by the dog than the more subjective breed
standards that are often difficult to operationalize. Furthermore, while an individual of a particular breed-
type is genetically predisposed to exhibit breed-typical behaviors, the proper rearing environment with
appropriate stimuli is also required for the expression of these behaviors (Coppinger and Coppinger,
2001). For example ,the unique stalking posture of the Border Collie is exhibited by both trained and
untrained individuals, but is refined by both training and experience (McConnell and Baylis, 1985). Given
that much of the empirical literature on breed differences assumes phylogenetic determinants of behavior,
the examination of breed-typical behaviors may represent a suitable direction for examining ontogenetic
factors that contribute to behavioral differences both between and within breeds.

In addition, differences in neurotransmitter levels (norepinephrine, dopamine, and epinephrine) have
been found between Border Collies, Siberian Huskies, and Shar Plaininetz that are consistent with
differences in their respective predatory motor patterns (Arons and Shoemaker, 1992). More recently,
differentially lower serotonin serum levels have been detected in aggressive English Cocker Spaniels
relative to aggressive dogs of other breeds (Amat et al., 2013), which may provide a neurological reason
for why this breed is often ranked as among the most likely to be seen by veterinarians and behaviorists
for human-directed aggression.

This is an important area for future research that will result in further scientific progress to understanding
breed differences in behavior.

9. Conclusions

Our review of primary literature, empirical data, and published veterinary and public safety records, suggests that
differences in behavior are evident among breeds of dogs. However, substantial within-breed differences inbehavior
also existeven in the most controlled experimental studies. Breed differences in behavior are therefore influenced by
both genetics, and by the environment and experience. Although breed and breed-group differences in behavior is
evident in many cases, there is little evidence that these behavioral differences correspond to conventional and
genetic categorizations of breeds. A more useful method of categorization may be based on directly observable breed-
typical behaviors, which are genetically inherited, tied to historical roles of the breed, and can be modified by
experience within an individuals lifetime. Different methodologies influence the results of studies, such that studies
using owner-based surveys are more likely to conclude breed differences exist than experimental studies, likely due
to the relatively small sample size and breed diversity in the latter. Interpretation of surveys conducted between
different countries should be done carefully. Behavioral scientists need to be aware of the limitations of working with
pet dogs with various rearing histories. Behavioral research is limited in its ability to make claims about breed
differences for three major reasons. First, definitive claims about breed differences in behavior cannot be made
without realizing that behavior is the outcome of complex interactions of phylogeny and ontogeny. Second, the large
number of breeds to be examined today simply discourages inquiry in behavioral research; and finally, the definitions
of many behavioral traits are subjective and structural, rather than objective and functional. It is the responsibility
not only of researchers in the field of canine behavior and genetics, but also of professionals who work with dogs,
including veterinarians, dog owners, breeders, obedience trainers and handlers of service dogs, and other parties
interested in the effectiveness of breed-specific legislation, to address the seissues if we are to further our
understanding of the many different dog breeds that exist among us in the world today.

Acknowledgments

We would like to thank Raymond P. Coppinger, Monique A.R. Udell, and four anonymous reviewers for their helpful
suggestions on an earlier version of this manuscript.

References

Amat, M., Brech, S.L., Camps, T., Torrente, C., Mariotti, V.M., Ruiz, J.L.,Manteca, X., 2013. Differences in serotonin
serum concentrationbetween aggressive English cocker spaniels and aggressive dogs ofother breeds. J. Vet. Behav. 8,
1925.

Amat, M., Manteca, X., Mariotti, V.M., de la Torre, J.L.R., Fatj, J., 2009. Aggressive behavior in the English cocker
spaniel. J. Vet. Behav. 4,111117.

American Kennel Club (AKC), 2006. The Complete Dog Book, 20th ed. Howell Book House, New York, NY.

Arons, C., Shoemaker, W., 1992. The distribution of catecholamines and ! b-endorphin in the brains of three
behaviorally distinct breeds of dogs and their F1 hybrids. Brain Res. 594, 3139.

Beaver, B., 1983. Clinical classification of canine aggression. Appl. Anim. Ethol. 10, 3543.

Beckmann, M., Enot, D.P., Overy, D.P., Scott, I.M., Jones, P.G., Allaway, D., Draper, J., 2010. Metabolite finger
printing of urine suggests breed-specific dietary metabolism differences in domestic dogs. Br. J. Nutr. 103, 1127
1138.

Beerda, B., Schilder, M., van Hooff, J., de Vries, H., Mol, J., 2000. Behaviouraland hormonal indicators of enduring
environmental stress in dogs. Anim. Welfare 9, 4962.

Blackwell, E.J., Bradshaw, J.W.S., Casey, R.A., 2013. Fear responses to noisesin domestic dogs: prevalence, risk
factors and co-occurrence with other fear related behaviour. Appl. Anim. Behav. Sci. 145, 1525.

Block, J., 1961. The Q-sort Method in Personality Assessment and Psychiatric Research. Charles C. Thomas,
Springfield, IL, 161 pp.

Bollen, K., Horowitz, J., 2008. Behavioral evaluation and demographic information in the assessment of
aggressiveness in shelter dogs. Appl. Anim. Behav. Sci. 112, 120135.

Borchelt, P., 1983. Aggressive behavior of dogs kept as companion animals: classification and influence of sex,
reproductive status and breed. Appl. Anim. Ethol. 10, 4561.

Bradley, J., 2011. The Relevance of Breed in Selecting a Companion Dog. National Canine Research Council, LCC,
United States.

Bradshaw, J.W.S., Goodwin, D., 1998. Determination of behavioural traitsof purebred dogs using factor analysis and
cluster analysis; a comparison of studies in the USA and UK. Res. Vet. Sci. 66, 7376.

Bradshaw, J.W.S., McPherson, J.A., Casey, R.A., Larter, I.S., 2002. Aetiology of separation-related behaviour in
domestic dogs. Vet. Rec. 151, 4346.

Brewer, D.J., Terence, C.S., Phillips, A., 2002. Dogs in Antiquity: Anubis to Cerbrus The Origins of the Domestic Dog.
Aris and Phillips, Warmin-ster, England.

Cattell, R., Bolz, C., Korth, B., 1973. Behavioral types in purebred dogs objectively determined by taxonome. Behav.
Genet. 3, 205216.

Christiansen, F., Bakken, M., Braastad, B., 2001. Behavioural differencesbetween three breed groups of hunting dogs
confronted with domestic sheep. Appl. Anim. Behav. Sci. 72, 115129.

Clutton-Brock, J., 1995. Origin of the dog: domestication and early history. In: Serpell, J.A. (Ed.), The Domestic Dog:
Its Evolution, Behaviour and Interactions with People. Cambridge University Press, Cambridge, pp.720.

Coppinger, R., Coppinger, L., 1996. Biologic bases of behavior of domestic dog breeds. In: Voith, V.L., Borchelt, P.L.
(Eds.), Readings in Companion Animal Behavior. Veterinary Learning Systems, Trenton, pp. 918.

Coppinger, R., Coppinger, L., 2001. Dogs: A New Understanding of Canine Origin. Behavior and Evolution, Scribner,
New York.

Coren, S., 1995. The Intelligence of Dogs: A Guide to the Thoughts, Emotions, and Inner Lives of Our Canine
Companions. Bantam Books, NewYork, NY.

Cornelissen, J.M., Hopster, H., 2010. Dog bites in The Netherlands: a study of victims, injuries, circumstances and
aggressors to support evaluation of breed specific legislation. Vet. J. 186, 277279.

Darwin, C., 1859. On the origin of species by means of natural selection. In: or the preservation of favoured races in
the struggle for life. Murray, London.

Diedrich, C., Giffroy, J., 2006. Behavioural testing in dogs: a review of methodology in search of standardization.
Appl. Anim. Behav. Sci. 97,5172.

Dorey, N., Udell, M., Wynne, C., 2009. Breed differences in dogs sensitivity to human points: a meta-analysis. Behav.
Processes 81, 409415.

Draper, T.W., 1995. Canine analogs of human personality factors. J. Gen.Psychol. 122, 241252.

Duffy, D., Hsu, Y., Serpell, J., 2008. Breed differences in canine aggression. Appl. Anim. Behav. Sci. 114, 441460.
University of Colorado Denver, 2010. Young, unsupervised children most at risk for dog bites, study shows: dogs
often target a childs face and eyes. Science Daily, Retrieved June 14, 2012, from
http://www.sciencedaily.com-/releases/2010/11/101109133233.htm.

Elliot, O., Scott, J., 1965. Analysis of breed differences in maze performancein dogs. Anim. Behav. 13, 518.

Ennik, I., Liinamo, A., Leighton, E., van Arendonk, J., 2006. Suitability forfield service in 4 breeds of guide dogs. J.
Vet. Behav. 1, 6774.

Fallani, G., Previde, E., Valsecchi, P., 2007. Behavioral and physiological responses of guide dogs to a situation of
emotional distress. Physiol. Behav. 90, 648655.

Fatj, J., Amat, M., Mariotti, V., de la Torre, J., Manteca, X., 2007. Analysis of 1040 cases of canine aggression in a
referral practice in Spain. J. Vet.Behav. 2, 158165.

Fox, M., 1964. The ontogeny of behaviour and neurologic responses in the dog. Anim. Behav. 12, 301310.

Freedman, D.G., 1958. Constitutional and environmental interaction inrearing of four breeds of dogs. Science 127,
585586.

Gcsi, M., Kara, E., Belenyi Bea Topal, J., Miklosi, A., 2009. The effect of development and individual differences in
pointing comprehension of dogs. Anim. Cognit. 12, 471479.

Gagnon, S., Dor, F., 1992. Search behavior in various breeds of adult dogs (Canis familiaris): object permanence
and olfactory cues. J. Comp. Psychol. 106, 5868.

Goddard, M., Beilharz, R., 1985a. A multivariate analysis of the genetics off earfulness in potential guide dogs. Behav.
Genet. 15, 6989.

Goddard, M., Beilharz, R., 1985b. Individual variation in agonistic behaviour in dogs. Anim. Behav. 33 (4), 1338
1342.

Goddard, M., Beilharz, R., 1983. Genetics of traits which determine the suitability of dogs as guide-dogs for the blind.
Appl. Anim. Ethol. 9,299315.

Goodwin, D., Bradshaw, J., Wickens, S., 1997. Paedomorphosis affects agonistic visual signals of domestic dogs.
Anim. Behav. 53,297304.

Green, J., Woodruff, R., 1983. The use of three breeds of dog to protect rangeland sheep from predators. Appl. Anim.
Ethol. 11,141161.

Guy, N., Luescher, U., Dohoo, S., Spangler, E., Miller, J., Dohoo, I., et al.,2001. Demographic and aggressive
characteristics of dogs in a general veterinary caseload. Appl. Anim. Behav. Sci. 74, 1528.

Harcourt, R.A., 1974. The dog in prehistoric and early historic Britain. J.Archaeol. Sci. 1, 151175.

Hart, B.L., Hart, L.A., 1985. Selecting pet dogs on the basis of cluster analysis of breed behavior profiles and gender.
J. Am. Vet. Med. Assoc. 186,11811185.

Hart, B.L., Miller, M.F., 1985. Behavioral profiles of dog breeds. J. Am. Vet.Med. Assoc. 186, 11751180.

Head, E., Callahan, H., Cummings, B., Cotman, C., Ruehl, W., Muggenberg, B., et al., 1997. Open field activity and
human interaction as a function of age and breed in dogs. Physiol. Behav. 62, 963971.

Helton, W., 2010. Does perceived trainability of dog (Canis lupus familiaris) breeds reflect differences in learning or
differences in physical ability? Behav. Processes 83, 315323.

Helton, W., Helton, N.D., 2010. Physical size matters in the domestic dogs (Canis lupus familiaris) ability to use
human pointing cues. Behav. Processes 85, 7779.

Hemmer, H., 1990. Domestication: The Decline of Environmental Appreciation. Cambridge University Press,
Cambridge, pp. 208.

Houpt, K.A., Willis, M.B., 2001. Genetics of behaviour. In: Ruvinsky, A., Sampson, J. (Eds.), The Genetics of the Dog.
CAB International, Wallingford, Oxon, pp. 371400.

Howell, T.J., Bennett, P.C., 2011. Puppy power! Using social cognition research tasks to improve socialization
practices for domestic dogs (Canis familiaris). J. Vet. Behav. 6, 195204.

Hsu, Y., Sun, L., 2010. Factors associated with aggressive responses in pet dogs. Appl. Anim. Behav. Sci. 123, 108
123.

Jakovcevic, A., Elgier, A., Mustaca, A., Bentosela, M., 2010. Breed differencesin dogs (Canis familiaris) gaze to the
human face. Behav. Processes 84,602607.

Jones, A., Gosling, S., 2005. Temperament and personality in dogs (Canis familiaris): a review and evaluation of past
research. Appl. Anim. Behav. Sci. 95, 153.

Kerswell, K., Butler, K., Bennett, P., Hemsworth, P., 2010. The relationships between morphological features and
social signaling behaviours injuvenile dogs: the effect of early experience with dogs of different morphotypes. Behav.
Processes 85, 17.

Kim, Y.K., Lee, S.S., Oh, S.I., Kim, J.S., Suy, E.H., Houpt, K.A., Lee, H.C., Lee ,H.J., Yeon, S.C., 2010. Behavioural
reactivity of the Korean native Jindodog varies with coat colour. Behav. Processes 84, 568572.

King, T., Hemsworth, P., Coleman, G., 2003. Fear of novel and start lingstimuli in domestic dogs. Appl. Anim. Behav.
Sci. 82, 4564.

Kobelt, A.J., Hemsworth, P.H., Barnett, J.L., Coleman, G.J., Butler, K.L., 2007. The behaviour of Labrador retrievers in
suburban backyards: the relationships between the backyard environment and dog behaviour. Appl. Anim. Behav.
Sci. 106, 7084. Landsberg, G.M., 1991. The distribution of canine behaviour cases at three behaviour referral
practices. Vet. Med. 86, 10811089.

Larson, G., Karlsson, E.K., Perri, A., et al., 2012. Rethinking dog domestication by integrating genetics, archeology,
and biogeography. Proc. Nat. Acad. Sci. 109, 88788883.

Ley, J., Bennett, P., Coleman, G., 2009. A refinement and validation of the Monash Canine Personality Questionnaire
(MCPQ). Appl. Anim. Behav. Sci. 116, 220227.

Ley, J., Coleman, G., Holmes, R., Hemsworth, P., 2007. Assessing fear of novel and startling stimuli in domestic dogs.
Appl. Anim. Behav. Sci. 104, 7184.

Lit, L., Schweitzer, J., Oberbauer, A., 2010. Characterization of humandog social interaction using owner report.
Behav. Processes 84, 721725.

Mahut, H., 1958. Breed differences in the dogs emotional behaviour. Can.J. Psychol. 12, 3544.

Marshall-Pescini, S., Passalacqua, C., Barnard, S., Valsecchi, P., Prato-Previde, E., 2009. Agility and search and
rescue training differently affects pet dogs behaviour in socio-cognitive tasks. Behav. Processes 81, 416422.

McConnell, P., Baylis, J., 1985. Interspecific communication in cooperative herding: acoustic and visual signals from
human shepherds and herding dogs. Z. Tierpsychol. 67, 302328.

McCrae, R.R., Costa Jr., P.T., Ostendorf, F., Angleitner, A., Hrebickova, M., Avia, M.D., Sanz, J., Sanchez-Bernardos,
M.L., Kusdil, M.E., Woodfield, R., Saunders, P.R., Smith, P.B., 2000. Nature over nurture: temperament, personality,
and life span development. J. Pers. Soc. Psychol. 78,173186.

McGreevy, P.D., Brueckner, A., Thomson, P.C., Branon, N.J., 2010. Motor laterality in four breeds of dog. J. Vet.
Behav. 5, 318323.

McKinley, J., Sambrook, T., 2000. Use of human-given cues by domestic dogs (Canis familiaris) and horses (Equus
caballus). Anim. Cognit. 3,1322.

Miklsi, A., 2007. Dog Behaviour, Evolution, and Cognition. Oxford University Press, New York, NY.

Mirk, E., Kubinyi, E., Gcsi, M., Miklsi, A., 2012. Preliminary analysis of an adjective-based dog personality
questionnaire developed to measure some aspects of personality in the domestic dog (Canis familiaris). Appl. Anim.
Behav. Sci. 138, 8898.

Morris, D., 2008. Dogs: The Ultimate Dictionary of Over 1,000 Dog Breeds. Trafalgar Square Publishing, North
Pomfret.

Nederhof, A.J., 1985. Methods of coping with social desirability bias: are view. Eur. J. Soc. Psychol. 15, 263280.

Notari, L., Goodwin, D., 2007. A survey of behavioural characteristics of purebred dogs in Italy. Appl. Anim. Behav.
Sci. 103, 118130.

Overall, K.L., Love, M., 2001. Dog bites to humansdemography, epidemiology, injury, and risk. J. Am. Vet. Med.
Assoc. 218, 19231934.

Passalacqua, C., Marshall-Pescini, S., Barnard, S., Lakatos, G., Valsecchi,P., Previde, E.P., 2011. Human-directed
gazing behaviour in puppies and adult dogs, Canis lupus familiaris. Anim. Behav. 82, 10431050.

Parker, H.G., Ostrander, E.A., 2005. Canine genomics and genetics: running with the pack. PLoS Genet. 1, 507513.

Parker, H.G., Shearin, A.L., Ostrander, E.A., 2010. Mans best friend becomes biologys best in show: genome
analyses in the domestic dog. Annu.Rev. Genet. 44, 309336.

Plutchik, R., 1971. Individual and breed differences in approach and with-drawal in dogs. Behaviour 40, 302311.

Podberscek, A.L., Serpell, J.A., 1996. The English cocker spaniel: preliminary findings on aggressive behavior. Appl.
Anim. Behav. Sci. 47,7589.

Pongracz, P., Miklosi, A., Vida, V., Csanyi, V., 2005. The pet dogs abilityfor learning from a human demonstrator in a
detour task is independent from the breed and age. Appl. Anim. Behav. Sci. 90 (3-4),309323.

Pullen, A.J., Merrill, R.J.N., Bradshaw, J.W.S., 2013. The effect of familiarity on behavior of kenneled dogs during
interactions with conspecifics. J. Appl. Anim. Welfare Sci. 16, 6476.

Reisner, I.R., Houpt, K.A., Frances, S.S., 1994. Differential diagnosis and management of human-directed aggression
in dogs. J. Am. Vet. Med.Assoc. 227, 15941603.

Riedel, J., Schumann, K., Kaminski, J., Call, J., Tomasello, M., 2008. The early ontogeny of humandog
communication. Anim. Behav. 75, 10031014.

Rine, 1991. Rosenthal, E. (Dec. 3, 1991) Study of Canine Genes Seeks Hintson Behavior. Times, Science, New York,
NY,
http://www.nytimes.com 1991/12/03/science/study-of-canine-genes-seeks-hints-on-behavior.html?pagewanted=all&src=pm
(retrieved on June 12,2012).

Ritvo, H., 1987. The Animal Estate: The English and Other Creatures in the Victorian Age. Harvard University Press,
Cambridge.

Riva, J., Marelli, S.P., Redaelli, V., Bondiolotti, G.P., Sforzini, E., Santoro, M.M., Carenzi, C., Verga, M., Luzi, F., 2012.
The effects of drug detection training on behavioral reactivity and blood neurotransmitter levels in drug detection
dogs: a preliminary study. J. Vet. Behav. 7,1120.

Roberts, T., McGreevy, P., Valenzuela, M., 2010. Human induced rotation and reorganization of the brain of domestic
dogs. PLoS One 5, e11946.Roll, A., Unshelm, J., 1997. Aggressive conflicts amongst dogs and factorsaffecting them.
Appl. Anim. Behav. Sci. 52, 229242.

Rooney, N., Bradshaw, J., 2004. Breed and sex differences in the behavioural attributes of specialist search dogsa
questionnaire survey of trainers and handlers. Appl. Anim. Behav. Sci. 86,123135.

Sacks, J., Sinclair, L., Gilchrist, J., Golab, G., Lockwood, R., 2000. Breeds of dogs involved in fatal human attacks in
the United States between 1979 and 1998. J. Am. Vet. Med. Assoc. 217, 836840.

Saetre, P., Strandberg, E., Sundgren, P., Pettersson, U., Jazin, E., Bergstrm,T., 2006. The genetic contribution to
canine personality. Genes Brain Behav. 5, 240248.

Schalke, E., Ott, S., Hirschfield, J., Hackbarth, H., 2010. Assessment of a Bull Terrier bloodline regarding possible
hypertrophic aggressive behaviour in situations of dogdog-contact of the temperament test of Lower Saxony. Berl.
Munch. Tierarztl. Wochenschr. 123, 192197.

Scott, J.P., Fuller, J.L., 1965. Genetics and the Social Behavior of the Dog. University of Chicago Press, Chicago, IL.

Serpell, J., 1995. The Domestic Dog: Its Evolution, Behaviour and Interaction with People. Cambridge University
Press, Cambridge.

Serpell, J.A., Hsu, Y., 2001. Development and validation of a novel method for evaluating behavior and temperament
in guide dogs. Appl. Anim.Behav. Sci. 72, 347364.

Serpell, J.A., Hsu, Y., 2005. Effects of breed, sex, and neuter status on trainability in dogs. Anthrozoos 18, 196207.

Sinn, D.L., Gosling, S.D., Hilliard, S., 2010. Personality and performance in military working dogs: reliability and
predictive validity of behavioral tests. Appl. Anim. Behav. Sci. 127, 5165.

Skinner, B.F., 1938. The Behavior of Organisms: An Experimental Analysis. Appleton-Century, Oxford, England.

Stafford, K.J., 1996. Opinions of veterinarians regarding aggression in different breeds of dogs. N.Z. Vet. J. 44, 138
141.

Svartberg, K., 2002. Shyness-boldness predicts performance in working dogs. Appl. Anim. Behav. Sci. 79, 157174.

Svartberg, K., Forkman, B., 2002. Personality traits in the domestic dog (Canis familiaris). Appl. Anim. Behav. Sci.
79, 133156.

Svartberg, K., 2006. Breed-typical behaviour in dogs-historical remnants or recent constructs? Appl. Anim. Behav.
Sci. 96, 293313.

Svartberg, K., 2007. Individual differences in behaviordog personality. In: Jensen, P. (Ed.), The Behavioural Biology
of Dogs. CAB International, Cambridge, pp. 182206.

Takeuchi, Y., Ogata, N., Houpt, K.A., Scarlett, J.M., 2001. Differences in background and outcome of three behavior
problems of dogs. Appl. Anim.Behav. Sci. 70, 297308.

Takeuchi, Y., Mori, Y., 2006. A comparison of the behavioral profiles of purebred dogs in Japan to profiles of those in
the United States and the United Kingdom. J. Vet. Med. Sci. 68, 789796.

Turcsn, B., Kubinyi, E., Miklsi, ., 2011. Trainability and boldnesstraits differ between dog breed clusters based on
conventional breed categories and genetic relatedness. Appl. Anim. Behav. Sci. 132,6170.

Udell, M.A.R., Ewald, M., Dorey, N.R., Wynne, C.D.L., 2014. Exploring breed differences in dogs (Canis lupus
familiaris): Does exaggeration or inhibition of predatory response predict performance on human-guided tasks? Anim.
Behav. 89, 99105.

Udell, M.A.R., Wynne, C.D.L., 2010. Ontogeny and phylogeny: both are essential to human-sensitive behavior in the
genus Canis. Anim. Behav. 79, e9e14.

Vs, J., Topl, J., Gcsi, M., Miklsi, ., Csnyi, V., 2005. A friend or anenemy? Dogs reaction to an unfamiliar person
showing behavioural cues of threat and friendliness at different times. Appl. Anim. Behav.Sci. 94, 99115.

Wayne, R.K., Ostrander, E.A., 2007. Lessons learned from the dog genome. Trends Genet. 23, 557567.

Werboff, J., Shepard, J., Scott, J., 1969. Transfer of training in dogs: effects of early training on similar and dissimilar
tasks in two breeds. J. Psychol.:Interdiscipl. Appl. 73, 121132.

Wilsson, E., Sundgren, P., 1997. The use of a behaviour test for the selection of dogs for service and breeding, I:
Method of testing and evaluating test results in the adult dog, demands on different kinds ofservice dogs, sex and
breed differences. Appl. Anim. Behav. Sci. 53,279295.

Wobber, V., Hare, B., Koler-Matznick, J., Wrangham, R., Tomasello, M., 2009. Breed differences in domestic dogs
(Canis familiaris) comprehension of human communicative signals. Interact. Stud.: Soc. Behav. Commun. Biol. Artif.
Syst. 10, 206224.

Wright, J., Nesselrote, M., 1987. Classification of behavior problems indogs: distributions of age, breed, sex and
reproductive status. Appl.Anim. Behav. Sci. 19, 169178.

Wright, J., Smith, A., Daniel, K., Adkins, K., 2007. Dog breed stereotype and exposure to negative behavior: effects
on perceptions of adoptability. J. Appl. Anim. Welfare Sci. 10, 255265.

Wynne, Clive D.L., 2007. What are animals? Why anthropomorphism is still not a scientific approach to behavior.
Comp. Cogn. Behav. Rev. 2,125135.

Xenophon, 1897. Three essays, on the duties of a calvery general, on horse-manship, and on hunting. In: The Works
of Xenophon by Xenophonand H.G. Dakyns, 2010. The MacMillan Company, New York, NY